[15:47:10] I was trying to do one for Swedish and got pretty far in the implementation in Z26783. But I am missing Item 4, which I can't get right. It should be similar to the iterable in map function, but I can't figure out what "wrapper" to choose to able to select it. Could someone help me with a clue? [16:03:28] Yeah, making one of those selectors is a bit of a mess right now. [18:28:08] When I click the 🛈 on a test to see the details it says in the gray box in the top "Something not working? Try Wikifunctions.Debug to trace your code." Is that also possible on a composition? [18:34:07] No… but the better news is Z26786. (re @Jan_ainali: When I click the 🛈 on a test to see the details it says in the gray box in the top "Something not working? Try Wikifunctions.Deb...) [18:36:51] Odd that the last test fails. I thought L1331846#S1-P5137 should have been enough. (re @Al: No… but the better news is Z26786.) [18:37:00] I re-used Z26671… two out of three tests pass. The third appears to be missing data. (re @Jan_ainali: I was trying to do one for Swedish and got pretty far in the implementation in Z26783. But I am missing an Item 4 in the concate...) [18:37:56] Is it not enough with a sense linking to the item? (re @Al: I re-used Z26671… two out of three tests pass. The third appears to be missing data.) [18:41:31] I think Z26671 uses the label from the Wikidata item but the gender from the lexeme. That needs thinking about! (re @Jan_ainali: Is it not enough with a sense linking to the item?) [18:42:31] Yeah, that is what I was trying to build in Z26783 but got stuck on the 4th item. (re @Al: I think Z26671 uses the label from the Wikidata item but the gender from the lexeme. That needs thinking about!) [18:43:43] I mean, essentially it is done in the third item, so it should be possible to be using a similar setup, but I have still not fully grasped how the types work so got stuck. [18:56:21] Yes… as Toby said, it’s an ugly workaround for a UI bug. It’s easier to add a Swedish label to Q47359 (assuming it’s missing). However… the more fundamental issue is that there is no guarantee that the text of the label uses a form of the same lexeme as the lexeme linked by sense. In the short term, we should have a Swedish Z22664 function that returns both the indefin [18:56:21] [18:56:22] ite article and the appropriate form of the noun, and that function ought to support the adjective being optional. (re @Jan_ainali: I mean, essentially it is done in the third item, so it should be possible to be using a similar setup, but I have still not ful...) [18:58:47] Oh yes, an indefinite noun phrase would be good to have! (re @Al: Yes… as Toby said, it’s an ugly workaround for a UI bug. It’s easier to add a Swedish label to Q47359 (assuming it’s missing). H...) [18:59:11] I'll get working on it. [19:25:02] I think I failed: Z26791 (re @Jan_ainali: I'll get working on it.) [20:39:07] It wasn’t far off, but I changed most of it. It will now return only a singular form and use “en” by default. We can easily switch that around if it’s not correct. The basic problem with the original implementation was comparing the first element of the list of genders with a list containing one gender. I don’t know when you would get more than one gender back for a Swe [20:39:07] [20:39:07] dish noun, but it will currently use “ett” if neuter is in the list. (re @Jan_ainali: I think I failed: Z26791) [20:42:57] Yeah, the order doesn't matter, and it should always only be one gender for Swedish nouns. (There really isn't a default, if it's not one, it's the other.) (re @Al: It wasn’t far off, but I changed most of it. It will now return only a singular form and use “en” by default. We can easily swit...) [20:49:29] So while this works, I still get stuck on that bug not letting me use a function call. : https://tools-static.wmflabs.org/bridgebot/255e3c70/file_72202.jpg [20:55:01] Yes. I can’t fix that, I’m afraid! 😟 (re @Jan_ainali: So while Z26791 works now, I still get stuck on that bug not letting me use a function call.) [22:17:21] Z26795 Here you can click the > chevron and replace the fetch function with the function you need to use instead (Z26791). (re @Jan_ainali: So while Z26791 works now, I still get stuck on that bug not letting me use a function call.)